MediaChance Dynamic Auto-Painter x64 Pro — Review

© 2011 Peter Free

 

05 December 2011

 

 

A lot less than meets one’s first impression

 

Lack of control and bland sameness is the name of the game with this software.

 

If you stay solely within Dynamic Auto-Painter x64 Pro — no matter what you do with its two relatively worthless masking functions — your output is going to look stylistically like everyone else’s (mostly bad) art.

 

See, for example, the representative Flickr.com gallery here.

 

Dynamic Auto-Painter is certainly very good at making photographs look like something else.  But that “something else” falls well short of what would legitimately qualify as being artful.

 

In order to get anything worthwhile from this software, I had to go back and forth to Photoshop — a process that was neither efficient nor artistically seamless.

 

 

Failure to follow through on a good idea

 

This software’s promise has been severely let down by its failure to follow through with obviously necessary refinements.

 

Auto-Painter may be worth its comparatively low price to people who are easily pleased and understand its limitations.

 

 

Specific criticisms

 

Poor documentation and nonexistent tutorials

 

Inability to layer

 

Inability to make “legitimate” masks

 

Generic “mushing” of critical elements

 

Randomly added small brush strokes have no connection to actual details

 

Sliders do significantly less than one would think they should

 

Too literal a way of thinking about artistic style

 

 

Poor documentation and nonexistent tutorials

 

Problems begin with the lack of tutorials and a decent instruction manual.  The PDF that is linked to the program is so elementary that it is near worthless.  And the software’s interface, though attractive, is similarly unhelpful.

 

Consequently, the only way to learn what Dynamic Auto-Painter can, or cannot, do is to use it for hours, while creating different permutations from the same photograph.

 

In my case, after accumulated days of use, I recognized that Auto-Painter really can’t do much of anything that I foresee regularly wasting time on.

 

 

Inability to layer

 

Layering is not an included function.

 

To layer, you’ll need to save whatever you made in Painter and take to another processing program for further alterations.  Then you’ll need to copy that result and bring it back into Painter.

 

The problem with this work-around is that Dynamic Auto-Painter was not designed according to Photoshop’s way of doing things.  For example, although you can create the equivalence of a mask (in Photoshop) by cutting out what you don’t want Auto Painter to see, as soon as you stick Photoshop’s see-through layer into Painter, Painter fills the “clear” portion of the layer with white.

 

That means that when you take Painter’s result back to Photoshop, you’ll have to cut the white off, so that you can continue to work with the layered Photoshop composite that you were attempting to build.

 

This leads to problems with sizing, too obvious artifacts, and transitions.  To compensate, you have to flatten the Photoshop composite, blend edges, paint, and do whatever else has become necessary as the result of moving back and forth between two philosophically different software programs.

 

 

Inability to make “legitimate” masks

 

Dynamic Auto-Painter’s masking deficiency is immediately apparent when trying to preserve facial details in an environmental portrait.

 

Ordinarily, when I think of masking, I think of a technique that allows me to prevent any manipulation of what I’m trying to protect.  Auto-Painter doesn’t agree.

 

Even when Painter’s portrait mask is applied, the software comes in with a sledgehammer to obliterate everything you wanted to preserve.  The portrait mask only prevents the most obnoxious of subsequently overlaid scribbling from worsening the situation.

 

 

Generic “mushing” of critical elements

 

Once you’ve worked with this software for a few days, you begin to see that its spits out a generic sameness within each of its alleged styles.

 

The software designer was apparently so determined to eliminate any semblance to a photograph that he or she created software that mushes everything up.  In fact, the program’s “Realistic” painting style is anything but.

 

Painter’s bland sameness appears to be the consequence of the program’s inability to actually follow the finer structure and details of the photographs that it destroys.  You can see this heavy-handed approach just as soon as the software begins the under-painting process.  Everything nuanced gets squashed.

 

 

Randomly added small brush strokes have no connection to actual details

 

After steamroller first and second phases, the program begins to sprinkle down an overlay of smaller strokes.  These are almost randomly laid and almost always miss the actual details in the original picture.  Therefore, they create a distracting mess in places where most users won’t want them.

 

More annoyingly, these added strokes are very much lighter or darker than those underneath.  They really stand out.  Lacking a connection to photograph’s own minutiae, Painter’s accent strokes are crassly intrusive.

 

There is no way to tell when the detail phase is going to kick in, without first running the picture through the transformation process.  If you do want the output to contain some small (yet visible) brush strokes — but you do not like what the program eventually does with its overlay of these — you will have to:

 

(a) remember when the software started doing what you did not want done,

 

(b) start over,

 

and

 

(c) stop the auto-painting process before it gets to the irritating part of the small brush phase.

 

The only good thing about this shortcoming is that the program counts its “brush strokes” during each phase of the transformation process.  If you can remember which brush phase the software was in at the time it went awry, the counter is helpful in estimating when to stop the new attempt.

 

Incidentally, MediaChance makes a big deal about its ability to control the application of small strokes in specific locations by clicking the mouse where you want these small strokes to go.  Given the comparative esthetic stupidity of the entire detail-adding process, the mouse click function is mostly useless.  I found that it ruined outcomes more often than it contributed anything of value.

 

 

Sliders do significantly less than one would think they should

 

The difference that (even extreme) adjustments of the control sliders make to output, in most cases, is comparatively trivial.  The resulting “painting” remains too visibly recognizable as having come from one of the program’s preset styles.

 

This is the same objection that I had to MediaChance’s HDR software during the two years that I used it.

 

 

Too literal a way of thinking about artistic style

 

One would think that emulating an artistic style would consist of generally imitating the chosen painter’s manipulation of paint and vision.  As opposed to dismembering a representative sample of that person’s work and transporting whole-hog bits of it into the stylized copy.

 

Dynamic Auto-Painter almost literally takes pieces of digitally copied paintings, cuts them up, and sticks these into its style generator.  So, let’s say you want to use either the Starry Night or Sunflower version of two of Auto-Painter’s Van Gogh presets.  You’re going to wind up with very visibly swirled stars or sunflowers in your output.

 

I’m not saying that the result necessarily looks bad, but — does emulating Van Gogh mean that we have to tolerate star and sunflower bits in something that has nothing to do with either?

 

Dynamic Auto-Painter’s generalized lack of color control presents an analogous problem.

 

The program assumes that its Van Gogh, Cézanne, Klimt, Monet, and “Camille” (a blend of Corot and Pissarro) presets all should come with a color palette that matches specific paintings these painters produced.

 

Consequently, Auto-Painter too often ignores the colors in your photograph and generates output that is coloristically unrelated to anything in the original scene.

 

You can partially adjust some of these “errors” in Photoshop.  But why should you have to?

 

 

Conclusion? — Lack of control and boring sameness argue against choosing Dynamic Auto-Painter

 

Whether Dynamic Auto-Painter is more than a “toy” is open to question.  It doesn’t give users enough control of its output to qualify as a serious artist’s daily tool.

 

But an easily pleased hobbyist will probably like it.  And I think I might occasionally use it in conjunction with Photoshop.

 

My chief criticism is that, without further modification, Dynamic Auto-Painter spits out the same bad art that we often see at lower tier craft shows.

 

If you have high standards, and are expecting a stand-alone program that will take you where your imagination directs, beware.